Volume 74, Issue 5, Pages 924-935 (June 2012) Neuronal Discharges and Gamma Oscillations Explicitly Reflect Visual Consciousness in the Lateral Prefrontal Cortex  Theofanis I. Panagiotaropoulos, Gustavo Deco, Vishal Kapoor, Nikos K. Logothetis  Neuron  Volume 74, Issue 5, Pages 924-935 (June 2012) DOI: 10.1016/j.neuron.2012.04.013 Copyright © 2012 Elsevier Inc. Terms and Conditions

Figure 1 Behavioral Task and Typical Single Unit Modulation (A) Physical alternation (upper panel) and BFS (lower panel) paradigm. For psychophysical measurements demonstrating the duration of perceptual suppression, see also Figure S1. (B) Example raster plots (in 20 ms time bins) and mean discharge responses for a single unit showing a strong preference for a monkey face during both physical alternation (left panel) and BFS (right panel). Stimuli insets depict the perceived visual pattern. Neuron 2012 74, 924-935DOI: (10.1016/j.neuron.2012.04.013) Copyright © 2012 Elsevier Inc. Terms and Conditions

Figure 2 Sensory versus Perceptual Modulation of Neuronal Discharges (A) Scatterplot of sensory versus perceptual preference (d′) for all 577 recorded single units. Grey dots represent single units showing no significant modulation in either of the two conditions. Green dots represent units showing significant modulation during both physical alternation and BFS. Red dots represent units that showed only significant sensory preference. Blue dots depict single units that exhibited significant modulation only during the BFS condition. Only 5% of the sensory modulated units (green dots, n = 3/64) fired more when their preferred stimulus was perceptually suppressed (represented here by a negative d′ value). The large majority of perceptually modulated single units followed phenomenal perception. For typical examples of perceptually modulated single units, see Figure S2. (B) Significant modulation during both physical alternation and BFS is more likely for units showing high sensory selectivity (d′ > 1). (C) Same as (A) for MUA. Here, each point represents the sensory versus perceptual selectivity (d′) of the summed, local spiking activity within a cortical site. The spiking activity in the large majority of sensory modulated sites followed the phenomenal perception of a preferred stimulus. (D) Same as (B) for MUA. Perceptual modulation is more likely for units showing stronger sensory selectivity (d′ > 1). Insets in bars in (B) and (D) show the ratio: number of sensory and perceptually modulated/number of sensory modulated cells. Neuron 2012 74, 924-935DOI: (10.1016/j.neuron.2012.04.013) Copyright © 2012 Elsevier Inc. Terms and Conditions

Figure 3 Mean SUA Population Response during Physical Alternation and BFS (A) Mean population SUA averaged across units showing significant sensory modulation during physical alternation. Blue curve depicts the mean SUA when visual stimulation starts from a nonpreferred stimulus followed by switching to a preferred visual pattern. Red curve depicts the mean SUA when visual stimulation starts with a preferred pattern followed by switching to a nonpreferred stimulus in the contralateral eye. (B) Same as (A) for BFS. Perceptual dominance of a preferred stimulus (blue, t = 1,001–2,000 ms) results in increased spiking activity similar to that observed during physical alternation of the same stimuli. When the same stimulus is perceptually suppressed (red, t = 1,001–2,000), the mean SUA remains suppressed. (C) Mean firing rate difference (nonpreferred − preferred) averaged across all significantly sensory modulated units during physical alternation (green curve) and BFS (orange curve) for the whole duration of visual stimulation. The mean firing rate difference was lower during BFS compared to physical alternation but still significantly higher than zero from t = 1,001–2,000 (running Wilcoxon signed-rank test, p < 0.05 for all time points examined). (D) Mean sensory (green curve) and perceptual (orange curve) ± SEM (thin lines) selectivity (d′) following a physical or perceptual stimulus transition. The magnitude of selectivity during BFS was high, albeit lower than the selectivity observed during physical alternation. Selectivity exhibited adaptation during the perceptual dominance of a preferred pattern but lasted for almost the whole second of visual competition. Firing rate difference (C) and d′ (D) were computed for 10 ms time bins. Neuron 2012 74, 924-935DOI: (10.1016/j.neuron.2012.04.013) Copyright © 2012 Elsevier Inc. Terms and Conditions

Figure 4 Mean MUA Population Response during Physical Alternation and BFS (A) Mean MUA population response during physical alternation, averaged across sites where spiking activity showed significant sensory modulation. Activity reflects monocular stimulus transitions between preferred and nonpreferred visual patterns (similar to Figure 3A). (B) Same as (A) for BFS. Modulation of local spiking activity is largely retained during subjective visual perception. (C) Mean firing rate difference (nonpreferred − preferred) during physical alternation (green curve) and BFS (orange curve) averaged across all sensory modulated units for the whole duration of visual stimulation. The mean firing rate difference was lower during BFS (orange) compared to physical alternation (green) but still significantly higher than zero (running Wilcoxon signed-rank test, p < 0.05 for all time points examined). (D) Sensory versus perceptual selectivity during the last second of visual stimulation. Perceptual modulation was reduced compared to physical alternation but remained above zero for most of the duration of visual competition. Firing rate difference (C) and d′ (D) were computed for 10 ms time bins. See also Figures S3 and S4. Neuron 2012 74, 924-935DOI: (10.1016/j.neuron.2012.04.013) Copyright © 2012 Elsevier Inc. Terms and Conditions

Figure 5 Mean LFP Modulation for All Recorded Sites with Significant Sensory Modulation (A) Mean power spectra for a period of 1,000 ms following a stimulus switch (t = 1,001–2,000 ms) across all sites showing significant MUA modulation during physical alternation. Note the dominant LPFC beta rhythm (15–35 Hz) appearing as a distinct peak in the power spectra. When a preferred (determined by using MUA modulation as a criterion) stimulus was perceived (blue) oscillatory power in the high frequency gamma range (here, 50–140 Hz) was higher compared to the mean power during the monocular presentation of a nonpreferred pattern (red). No effect is observed in the beta band. Spectral power peak in 50 Hz reflects power line noise. (B) Same as (A) for BFS. Oscillatory power in the high-frequency range was increased when a preferred stimulus was perceptually dominant (blue) compared to the respective power when the same stimulus was perceptually suppressed (red). LFP power in frequencies between 15 and 30 Hz decreased when a preferred (by the MUA and high-frequency LFP power) stimulus was perceptually dominant compared to the power measured during its perceptual suppression. Insets in (A) and (B) are magnified plots of the high-frequency spectral differences. (C) Quantification of the mean power difference, observed in (A) and (B), between preferred and nonpreferred stimulus for physical alternation (green curve, mean [thick lines] ± SEM [thin lines]) and BFS (orange curve, mean [thick lines] ± SEM [thin lines]). For frequencies higher than 50 Hz, oscillatory power was higher when a preferred stimulus was perceived without significant difference between purely sensory stimulation and BFS. In frequencies between 15 and 30 Hz, there is a trend of oscillatory power to decrease during the perceptual dominance of a preferred stimulus in BFS (black arrow). (D) Same as (C), using the d′ as a measure of LFP power modulation. Similar to (C), there are no remarkable differences in the modulation of high frequencies between physical alternation and BFS. Intermediate frequencies (15–30 Hz) decrease during the perceptual dominance of a preferred stimulus (black arrow). See also Figure S5. Neuron 2012 74, 924-935DOI: (10.1016/j.neuron.2012.04.013) Copyright © 2012 Elsevier Inc. Terms and Conditions

Figure 6 Time-Frequency LFP Modulation (A) Time-frequency plot during physical alternation, averaged across the population of significantly sensory modulated sites. The average (across trials) spectrogram of each recording site for preferred to nonpreferred transitions was subtracted from the respective spectrogram obtained for nonpreferred to preferred transitions. (B) Same as (A) during BFS. High-frequency modulation was retained during almost one second of subjective visual perception. In addition, 15–30 Hz power slightly decreased during the perceptual dominance of a preferred pattern specifically during BFS, as indicated by arrows in (A) and (B). See also Figure S6. Neuron 2012 74, 924-935DOI: (10.1016/j.neuron.2012.04.013) Copyright © 2012 Elsevier Inc. Terms and Conditions